mljar-supervised
Python package for AutoML on Tabular Data with Feature Engineering, Hyper-Parameters Tuning, Explanations and Automatic Documentation
What is mljar-supervised?
Python package for AutoML on Tabular Data with Feature Engineering, Hyper-Parameters Tuning, Explanations and Automatic Documentation
Overview of mljar-supervised
What it is
mljar-supervised is a Python package for automated machine learning (AutoML) on tabular data. It trains models, discovers features, and explains predictions. The package supports classic machine learning algorithms, new neural architectures, and tabular foundation models through the same Python interface.
Key features
- Automatic model training with comparison across multiple algorithms, including linear models, Random Forest, XGBoost, LightGBM, CatBoost, RealMLP, TabM mini, Kumo, and TabICLv2
- Automatic documentation: every run creates a leaderboard and individual model reports with feature importance, learning curves, confusion matrices, and SHAP explanations
- Feature engineering: handles missing values and categorical inputs, can construct numeric combinations with golden features, and can use feature discovery to propose additional columns
- Fairness aware training: measures group differences and can train reweighted candidates to try to meet a selected threshold
- Optional support for neural architectures (RealMLP, TabM) and tabular foundation models (Kumo, TabICLv2) via package extras
Use cases
This package is designed for tabular data tasks such as classification and regression. It can be used for customer churn prediction, concrete strength estimation, and other supervised learning problems where automated model selection, feature engineering, and explanation are needed.
Getting started
Use Python 3.11 or newer. In a virtual environment, install the release candidate:python -m pip install 'mljar-supervised==1.4.0rc6'See the repository README for full installation and usage instructions.
License and pricing
This package is open source under the MIT license. Pricing is not stated; the project is listed as open source.
Based on the project's public README and repository metadata.